Background: Radiofrequency ablation (RFA) has gained ground as an effective and well-tolerated technique to treat benign thyroid nodules. Most of the available studies have described the short-term outcomes of RFA, whereas there is a limited number of studies evaluating long-term issues, such as regrowth and the likelihood of retreatments. In addition, risk markers of regrowth and retreatment remain to be defined. The initial ablation ratio (IAR) is an index that measures the amount of ablation after RFA, which has been associated with technique efficacy (i.e. volume reduction >50% after 1 year from the procedure). This study aimed at evaluating i) IAR reproducibility and ii) IAR predictive value for RFA 5-year outcomes.Materials and Methods: This is a retrospective single center study on patients with benign thyroid nodules treated with RFA and followed for 5 years after initial treatment. IAR interobserver reproducibility was evaluated with Bland-Altman method and Lin's concordance correlation coefficient (rhoc). IAR predictive value for RFA 5-year outcomes was evaluated with linear and logistic regression models, as well as with Cox models, while receiver operating characteristic (ROC) analyses were used for cut-offs.Results: We selected 78 patients with 82 benign thyroid nodules. The procedure significantly reduced nodule volume and this reduction was generally maintained over time. Technique efficacy was achieved in 92% of patients, while 23% of nodules regrew and 12% of nodules were retreated. Median IAR was 83%. Lin's concordance and Pearson's correlation coefficients suggested a good interobserver reproducibility of this index, consistent with the limits of agreement of the Bland-Altman plot. IAR was significantly associated with technique efficacy, 1- and 5-year volume reduction ratio, and with the likelihood of a retreatment, but not with nodule regrowth. ROC analyses showed that IAR cut-off was 49% for technique efficacy and 73% for retreatment.Conclusions: Our results show for the first time that IAR is reproducible and that it predicts the volume reduction and the likelihood of a retreatment after 5 years from RFA.
